At Katherine, N.T., stocking rates of 1.33, 1.00, and 0.67 Shorthorn steers an acre for a period of 18 weeks in the wet season beginning January were imposed on established pastures sown to Townsville sty10 (Stylosanthes humilis) alone but subsequently invaded to varying degrees by annual grasses. The experiment was repeated over three years. There were no significant differences in mean liveweight gain per head over three years ; hence liveweight gain per acre at the heaviest stocking rate was approximately twice that at the lightest rate. The mean dry matter yield of grass increased significantly with decreasing stocking rate, but there were no significant differences between stocking rates in Townsville stylo dry matter yield, nitrogen yield, or phosphorus yield. Correlation coefficients were calculated using data from individual paddocks and years. There was a highly significant positive correlation between liveweight gain per head and pasture nitrogen yield, and between liveweight gain per head and pasture phosphorus yield.